Data Visualization Assignment

Visualization 1

Interpretation

This data depicts a subset of approx. 209,000 yelp businesses in ten metropolitan areas within the US and Canada. As can be seen in the figure businesses that accept creditcards, usually have more reviews especially in the coastal areas in the West and the East.

Discussion

Discussion of key findings: The dataset depicts that businesses are rather ready to accept creditcards and have, on average, more ratings than those who don't. Based on the colors it appears debateable if this results in a better rating. At least it appears as though this factor has no significant effect on the rating. However questions remain whether this dataset is representative. Also questions have to be raised regarding a hidden variable. Maybe the higher amount of ratings is due to the location of the restaurant or any other factor. Here statistical testing and correlation analysis would be necessary.

Challenges encountered:

• Since I wanted to analyze a business sample about life in America a lot of platforms and databases were available. However acessing those databases via download or API was difficult as some had denied restrictions or hard to operate.
• The dataset consisted of multiple JSON files. However due to sheer size of the data, computation was difficult and consumed valuable time. It didn't help that memory and CPU were operating close to maximum capacity all the time.
• The dataset consisted of a multiple dictionaries inside a column which was difficult to extract. Since the dictionary was inside of quotation marks the extraction proved to be difficult and memory consuming.
• Also the list contained nan values which had to be interpreted. Therefore, I assumed nan values to be businesses not accepting credit cards.
• The integration of a big and dynamic image was slowing down the process of loading the HTML webpage.

Sources

• Find interesting public data: https://rs.io/100-interesting-data-sets-for-statistics and https://www.forbes.com/sites/bernardmarr/2018/02/26/big-data-and-ai-30-amazing-and-free-public-data-sources-for-2018/ and https://www.springboard.com/blog/free-public-data-sets-data-science-project/
• Look at different datasets not picked: https://www.bea.gov/classroom and http://millionsongdataset.com/lastfm/ and https://apps.ecmwf.int/datasets/ and http://insideairbnb.com/get-the-data.html and https://data.census.gov/cedsci/ andhttps://azure.microsoft.com/en-us/services/open-datasets/catalog/us-labor-force-statistics/
• Dataset chosen:https://www.yelp.com/dataset
• Import JSON file: https://stackoverflow.com/questions/30088006/loading-a-file-with-more-than-one-line-of-json-into-pandas
• Import JSON file 2: https://www.dataquest.io/blog/python-json-tutorial/
• Convert string dict to regular dict: https://www.tutorialspoint.com/How-to-convert-a-String-representation-of-a-Dictionary-to-a-dictionary-in-Python
• Adapt geoscatter properties: https://plotly.com/python-api-reference/generated/plotly.express.scatter_geo.html
• Throubleshoot geoscatter: https://stackoverflow.com/questions/52771328/plotly-chart-not-showing-in-jupyter-notebook
• Embed geoscatter into HTML website: https://matteoguzzo.com/blog/embed-html-graphs-plotly/ and https://stackoverflow.com/questions/36262748/python-save-plotly-plot-to-local-file-and-insert-into-html

About Me